1. Moz and Retinoic Acid Coordinately Regulate H3K9 Acetylation, Hox Gene Expression, and Segment Identity 
Anne K. Voss, Caitlin Collin, Mathew P. Dixon, Tim Thomas 
Developmental Cell 
Volume 17, Issue 5, Pages (November 2009)
DOI: /j.devcel
Copyright © 2009 Elsevier Inc. Terms and Conditions

2. Figure 1
MozΔ/Δ Pups Have an Anterior Homeotic Transformation of the Axial Skeleton and the Nervous System
E18.5 (A–H) and E10.5 (I–L) wild-type (A, C, E, G, I, and K) and MozΔ/Δ (B, D, F, H, J, and L). Shown are the external appearance (A and B) and skeletal preparations of the cervical vertebrae column: lateral view (C and D), dorsal view (E and F), and axial, rostral view of individual cervical vertebra (G and H). Neurofilament immunostaining displays the segmental nerves along the body axis (J–L). Note the hunched posture and elongated neck region (bracket in [B] versus [A]). MozΔ/Δ mutant pups have eight instead of seven cervical vertebrae (C–H). The additional cervical vertebra is most similar to the morphology of the atlas (C, D, G, and H). The first and the supernumerary atlas share one enlarged anterior arch of the atlas (solid arrows in [D] versus [C]). The identities of all other cervical vertebrae are shifted to anterior by one segment (G and H). The tuberculum anterior is absent in the MozΔ/Δ (stippled arrow in [C] and [D]). In some MozΔ/Δ pups additional arches or partial arches of supernumerary atlases can be observed (asterisk in [F] versus [E]). The first nerve contributing to the innervation of the forelimb is the 7th intersegmental nerve in the wild-type and the 8th in the MozΔ/Δ (arrowheads in [J] versus [I]). The first five intersegmental nerves contribute to the hypoglossal nerve (XII) in MozΔ/Δ as compared to the first four in wild-type (arrowheads in [L] versus [K]). I–XII, 1st to 12th cranial nerve; 1–8, 1st to 8th intersegmental nerve; aaa, anterior arch of the atlas; At, Atlas; At', supernumerary atlas; Ax, axis; C3–C7, 3rd to 7th cervical vertebrae; Ax', C4′–C8′, axis and C4 to C8 in the MozΔ/Δ (prime indicates that these differ from the wild-type vertebrae); Eo, exooccipital bone; FB, forebrain; FL, forelimb; HB, hyoid bone; He, heart; PA1, PA2, 1st, 2nd pharyngeal arch; R1–R3, 1st to 3rd rib; Rh2–Rh7, rhombomeres 2 to 7; T1, T1′, 1st thoracic vertebra; ta, tuberculum anterior; TR, tympanic ring; S1, 1st bone segment of the sternum; So, supraoccipital bone. Scale bar equals 2.5 mm in (A) and (B), 1.9 mm in (C) and (D), 0.6 mm in (E) and (F), 1.3 mm (G) and (H), 500 μm in (I) and (J), and 300 μm in (K) and (L).
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

3. Figure 2
MozΔ/Δ Show Reduced Expression of all Hoxa Cluster Genes and a Posterior Shift in Hoxa Gene Expression from the 3rd Gene Onward
Wild-type (A, C, E, G, I, and K) and MozΔ/Δ (B, D, F, H, J, and L) at E8.5 (A and B) and E10.5 (C–L). In situ hybridization detected Hoxa1 to Hoxa6 mRNA as indicated. E10.5 sense control experiments did not produce a signal (data not shown). Note the weaker signal in the MozΔ/Δ, in particular Hoxa1 (arrows [B] versus [A]) and Hoxa3 ([F] versus [E]). The anterior expression boundary in the neural tube is indicated (arrows in [E] and [F], dotted lines in [I]–[L], distance from the otic vesicle in [G] and [H]). Very weak expression of Hoxa2 in the 3rd rhombomere in MozΔ/Δ is indicated (R3 in D versus C). 1a, 1b, maxillary and mandibular part of the 1st pharyngeal arch; 2, 3, 2nd and 3rd pharyngeal arch; •, prevertebrae; ∗, otic vesicle. Scale bars equal 550 μm in (A) and (B), 455 μm in (C), (D), (K) and (L), 755 μm in (E) and (F), 1.2 mm in (G) and (H), and 485 μm in (I) and (J).
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

4. Figure 3
MozΔ/Δ Show Reduced Expression of All Hoxb Cluster Genes and a Posterior Shift in Hoxb Gene Expression from the 3rd Gene Onward
E10.5 wild-type (A, C, E, G, I, K, and M) and MozΔ/Δ (B, D, F, H, J, L, and N). In situ hybridization (A–L) detecting Hoxb2 to Hoxb6 and Hoxb8 mRNA as indicated. Note the weaker signal in the MozΔ/Δ, in particular Hoxb3 (C and D) and Hoxb4 (F] versus [E]). Hoxb4 protein distribution is shown (M and N). The anterior expression boundary in the neural tube is indicated (arrows in [A]–[D], dotted lines in [E], [F], and [I]–[L], and distance from the otic vesicle in [G], [H], [M], and [N]). Labeling is as in Figure 2. Scale bar equals 1.1 mm in [A]–[D], 390 μm in [E], [F], [M], and [N], 725 μm in [G] and [H], 340 μm in [I] and [J], and 390 μm in [K] and [L].
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

5. Figure 4
Hox Gene Expression in the MozΔ/Δ Assessed by Whole-Mount In Situ Hybridization and Northern Blot
(A) Diagram of the developing axial skeleton from the 1st occipital to the 9th thoracic segment. Bars show the anterior extent of Hox gene expression domains in the paraxial mesoderm at E10.5. Data are shown of a total of 60 wild-type and 60 MozΔ/Δ embryos, 6 of each genotype for each probe. Note the shift by at least one segment posterior from the 3rd paralogous group onward (Hoxa3, Hoxb3, and more posteriorily expressed Hox genes). Shaded area indicates the first segment, which lost expression of both paralogous genes of Hox clusters A and B. The identity of this segment is duly shifted anteriorily (C2→C1).
(B–D) Quantitation of the reduction in Hox gene expression in MozΔ/Δ and controls by northern blot analysis of pooled E9.5 and E10.5 embryos (B) and of individual E11.5 embryos followed by densitometry (C and D) shows an approximately 2-fold reduction in expression of Hoxa3, Hoxa4, Hoxb3, and Hoxb4 (n = 3 to 5 E11.5 embryos per genotype; p < ). Data are presented and were analyzed as described in the Experimental Procedures. Error bars show SEM. p values of <0.01 and <0.001 are indicated with two and three asterisks, respectively. At E7.5 Hoxa3 expression was very low or undetectable in Moz−/− mutants as compared to controls ([E] and [F] versus [H] and [I]) and did not reach the same extent and intensity (G and J). PS, primitive streak; asterisks in (H) and (K) indicate somites; the arrow indicates the hindbrain expression domain of Hoxa3 detectable in the wild-type. Bars equal 180 μm in (F), (G), (I), and (J) and 200 μm in (H).
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

6. Figure 5
H3K9 Hypoacetylation and Hypermethylation at Hox Loci in Moz−/− Embryos
(A) Genomic region examined for histone modifications and protein binding by ChIP/qPCR. “Hoxa3_1”, “Hoxa4_1”, “_2,” etc denote sequences amplified in (B)–(M) and Figure 7. H3K9ac (B–E), H3K9me3 (H and I) and H3K14ac (J–M) were assessed by ChIP of lysates of Moz−/− mutant and wild-type E10.5 embryos followed by qPCR of sequences spanning the start of transcription ([B, H, and J], tile_1), 0.7 to 1.5 kb 5′ of the start of transcription ([C, H, and K], tiles_2 and _3), in the protein coding regions ([D and L], cds), and at putative RARE (E and M). Data are derived from a total of 30 embryos, 6 of each genotype for H3K9ac and H3K14ac, 3 of each genotype for H3K9me3. Note the reduction in H3K9ac in Moz−/− mutant embryos at Hox loci overall ([B–E], overall p < ) and the corresponding increase in H3K9me3 ([H and I], overall p = ), but the lack of an effect on H3K14ac ([J–M], overall p = ). (F and G) Global H3K9 acetylation as assessed by immunoblotting (F) and densitometry (G) was not significantly different in the three genotypes (Mr 17 kDA; p = ; n = 3 E10.5 embryos per genotype; 1 individual embryo per lane). Data are presented and were analyzed as described in the Experimental Procedures. Error bars represent SEM. For the effect of Moz genotype, p values <0.05, <0.01, and <0.001 are indicated with one, two, and three asterisks, respectively. Exact p values are given in Table S2.
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

7. Figure 6
Retinoic Acid Treatment Rescues the MozΔ/Δ Mutant Body Segment Identity Defect
Skeletal preparations of wild-type (A and C) and MozΔ/Δ (B and D) E18.5 pups. Dams were treated at E7.25 with peanut oil (A and B) or 10 μg/g bodyweight RA in peanut oil (RA, [C and D]). Note the supernumerary cervical vertebra in the peanut-oil-treated MozΔ/Δ ([B] versus [A]), as seen in untreated pups (Figure 1), and the rescue of the cervical vertebrae in the RA-treated MozΔ/Δ ([D] versus [C] and [A]), as well as the reduction of the number of cervical vertebrae to six in the RA-treated wild-type ([C] versus [A]). (E) Enumeration of skeletal preparations. The effects of RA on MozΔ/Δ (rescue), on wild-type (reduction to six cervical vertebrae), and on MozΔ/+ heterozygote (reduction to six cervical vertebrae in half the cases) was statistically significant with p < , p < , and p = 0.009, respectively. Data were analyzed as described in the Experimental Procedures. Labeling as in Figure 1. The scale bar equals 0.6 mm.
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions

8. Figure 7
Retinoic Acid Treatment of Moz−/− Mutants at E9.5 Restores H3K9 Acetylation at E10.5, and Recruitment of Ing5 and Mll1 Are Dependent on Moz
E10.5 embryos were used for ChIP experiments (A–P) as in Figures 5O and 5P) or after treatment of the pregnant females with vehicle (peanut oil) or RA at E9.5. H3K9 (A–E), H3K14 (F–H), and H4K16 (I–K) acetylation and Ing5 (L–N) and Mll1 association (O and P) were assessed at a subset of the genomic Hox sites examined in Figure 5. Effects summarized over all genomic sites are displayed (E, H, K, N, and P). Note the reduction in H3K9ac in Moz−/− mutants without RA as compared to controls (A, B, and E) and the similar levels of H3K9ac after RA treatment (C, D, and E), whereas H4K16ac is increased in Moz−/− mutants and not affected by RA treatment (I–K). In contrast to H3K9ac and H4K16ac, H3K14ac is generally not different between Moz−/− mutants and controls irrespective of RA (F–H). Note the very low levels of Ing5 association with Hox loci in Moz−/− mutants as compared to controls (L and N). RA treatment did not rescue the levels of Ing5 at Hox loci in Moz−/− mutants (M and N). Note the reduction in the TrxG protein Mll1 at Hox loci in Moz−/− mutants as compared to controls (O and P). For the effect of Moz genotype, p values are indicated as in Figure 5, and exact p values are in Table S3. Error bars represent SEM.
Developmental Cell  , DOI: ( /j.devcel )
Copyright © 2009 Elsevier Inc. Terms and Conditions